- 11 Feb, 2007 40 commits
-
-
Jiri Slaby authored
2 spin_unlocks are omitted in the interrupt handler. Put them there to fix up deadlocking on UP. Signed-off-by:
Jiri Slaby <jirislaby@gmail.com> Signed-off-by:
Andrew Morton <akpm@linux-foundation.org> Signed-off-by:
Linus Torvalds <torvalds@linux-foundation.org>
-
Adrian Bunk authored
Signed-off-by:
Adrian Bunk <bunk@stusta.de> Cc: Jiri Slaby <jirislaby@gmail.com> Signed-off-by:
-
Andreas Jaggi authored
Removes an unused and ambiguous redefinition of INIT_WORK() Signed-off-by:
Andreas Jaggi <andreas.jaggi@waterwave.ch> Acked-by:
-
Eric Dumazet authored
I added IS_NOATIME(inode) macro definition in include/linux/fs.h, true if the inode superblock is marked readonly or noatime. This new macro is then used in touch_atime() instead of separatly testing MS_RDONLY and MS_NOATIME Signed-off-by:
Eric Dumazet <dada1@cosmosbay.com> Signed-off-by:
-
Eric Dumazet authored
I noticed cache misses in touch_atime() that can be avoided if we keep mnt_count & mnt_expiry_mark in a different cache line than mnt_flags (mostly read) mnt_count & mnt_expiry_mark are modified each time a file is opened/closed in a file system. touch_atime() is called each time a file is read, and generally needs to read mnt_flags. Other fields of struct vfsmount are mostly read so I chose to move mnt_count & mnt_expiry_mark at the end of struct vfsmount. And adding a comment so that nobody tries to re-arrange fields to fill the holes :) On 64bits platforms, the new offsetof(mnt_count) is 0xC0 On 32bits platforms, it is 0x60, so I didnot add a ____cacheline_aligned_in_smp because it would have a too big impact on the size of this object (in particular if CONFIG_X86_L1_CACHE_SHIFT=7) Signed-off-by:
-
Jiri Slaby authored
- tty_hangup() itself schedules work, so there is no need to schedule hangup in the driver - tty_wakeup(): it's safe to call it while in atomic, so that its schedule_work might be also wiped out Signed-off-by:
Alan Cox <alan@redhat.com> Signed-off-by:
-
Brent Casavant authored
The SGI IOC3 and IOC4 PCI devices implement memory space apertures, not I/O space apertures. Use the appropriate region management functions. Signed-off-by:
Brent Casavant <bcasavan@sgi.com> Cc: Pat Gefre <pfg@sgi.com> Cc: Stanislaw Skowronek <skylark@linux-mips.org> Cc: Brent Casavant <bcasavan@sgi.com> Acked-by:
Ralf Baechle <ralf@linux-mips.org> Signed-off-by:
-
Luciano Rocha authored
Extend usr/gen_init_cpio.c "file" entry, adding support for hard links. Previous format: file <name> <location> <mode> <uid> <gid> New format: file <name> <location> <mode> <uid> <gid> [<hard links>] The hard links specification is optional, keeping the previous behaviour. All hard links are defined sequentially in the resulting cpio and the file data is present only in the last link. This is the behaviour of GNU's cpio and is supported by the kernel initramfs extractor. Signed-off-by:
Luciano Rocha <strange@nsk.no-ip.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by:
-
Jean-Paul Saman authored
Update all arch/*/kernel/vmlinux.lds.S to not include space for initramfs when CONFIG_BLK_DEV_INITRAMFS is not selected. This saves another 4 kbytes on most platfoms (some reserve PAGE_SIZE for initramfs). Signed-off-by:
Jean-Paul Saman <jean-paul.saman@nxp.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: <linux-arch@vger.kernel.org> Signed-off-by:
-
Jean-Paul Saman authored
The file init/initramfs.c is always compiled and linked in the kernel vmlinux even when BLK_DEV_RAM and BLK_DEV_INITRD are disabled and the system isn't using any form of an initramfs or initrd. In this situation the code is only used to unpack a (static) default initial rootfilesystem. The current init/initramfs.c code. usr/initramfs_data.o compiles to a size of ~15 kbytes. Disabling BLK_DEV_RAM and BLK_DEV_INTRD shrinks the kernel code size with ~60 Kbytes. This patch avoids compiling in the code and data for initramfs support if CONFIG_BLK_DEV_INITRD is not defined. Instead of the initramfs code and data it uses a small routine in init/noinitramfs.c to setup an initial static default environment for mounting a rootfilesystem later on in the kernel initialisation process. The new code is: 164 bytes of size. The patch is separated in two parts: 1) doesn't compile initramfs code when CONFIG_BLK_DEV_INITRD is not set 2) changing all plaforms vmlinux.lds.S files to not reserve an area of PAGE_SIZE when CONFIG_BLK_DEV_INITRD is not set. [deweerdt@free.fr: warning fix] Signed-off-by:
Frederik Deweerdt <frederik.deweerdt@gmail.com> Signed-off-by:
-
Jiri Slaby authored
tty_wakeup calls wake_up_interruptible(&tty->write_wait) itself, it's not needed to wake up again after tty_wakeup returns. Signed-off-by:
-
Andrew Morton authored
- Prevent things like this: block/ll_rw_blk.c: In function 'submit_bio': block/ll_rw_blk.c:3222: warning: unused variable 'count' inlines are very, very preferable to macros. - remove unused get_cpu_vm_events() macro Cc: Christoph Lameter <clameter@engr.sgi.com> Signed-off-by:
-
Adrian Bunk authored
include/linux/byteorder/pdp_endian.h is completely unused, and the comment in the file itself states that it's both untested and only a proof-of-concept. Signed-off-by:
-
Alexey Dobriyan authored
currently it's 1) if *oldlenp == 0, don't writeback anything 2) if *oldlenp >= table->maxlen, don't writeback more than table->maxlen bytes and rewrite *oldlenp don't look at underlying type granularity 3) if 0 < *oldlenp < table->maxlen, *cough* string sysctls don't writeback more than *oldlenp bytes. OK, that's because sizeof(char) == 1 int sysctls writeback anything in (0, table->maxlen] range Though accept integers divisible by sizeof(int) for writing. sysctl_jiffies and sysctl_ms_jiffies don't writeback anything but sizeof(int), which violates 1) and 2). So, make sysctl_jiffies and sysctl_ms_jiffies accept a) *oldlenp == 0, not doing writeback b) *oldlenp >= sizeof(int), writing one integer. -EINVAL still returned for *oldlenp == 1, 2, 3. Signed-off-by:
Alexey Dobriyan <adobriyan@openvz.org> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by:
-
Eric W. Biederman authored
This does several things. - It moves looking up of the current foreground console into process context where we can safely take the semaphore that protects this operation. - It uses the new flavor of work queue processing. - This generates a factor of do_SAK, __do_SAK that runs immediately. - This calls __do_SAK with the console semaphore held ensuring nothing else happens to the console while we process the SAK operation. - With the console SAK processing moved into process context this patch removes the xchg operations that I used to attempt to attomically update struct pid, because of the strange locking used in the SAK processing. With SAK using the normal console semaphore nothing special is needed. Cc: Oleg Nesterov <oleg@tv-sign.ru> Signed-off-by:
Eric W. Biederman <ebiederm@xmission.com> Signed-off-by:
-
Michael Neuling authored
Add retain_initrd option to control freeing of initrd memory after extraction. By default, free memory as previously. The first boot will need to hold a copy of the in memory fs for the second boot. This image can be large (much larger than the kernel), hence we can save time when the memory loader is slow. Also, it reduces the memory footprint while extracting the first boot since you don't need another copy of the fs. Signed-off-by:
Michael Neuling <mikey@neuling.org> Cc: "Randy.Dunlap" <rdunlap@xenotime.net> Signed-off-by:
-
Miguel Ojeda Sandonis authored
Add support for auxiliary displays, the ks0108 LCD controller, the cfag12864b LCD and adds a framebuffer device: cfag12864bfb. - Add a "auxdisplay/" folder in "drivers/" for auxiliary display drivers. - Add support for the ks0108 LCD Controller as a device driver. (uses parport interface) - Add support for the cfag12864b LCD as a device driver. (uses ks0108 LCD Controller driver) - Add a framebuffer device called cfag12864bfb. (uses cfag12864b LCD driver) - Add the usual Documentation, includes, Makefiles, Kconfigs, MAINTAINERS, CREDITS... - Miguel Ojeda will maintain all the stuff above. [rdunlap@xenotime.net: workqueue fixups] [akpm@osdl.org: kconfig fix] Signed-off-by:
Miguel Ojeda Sandonis <maxextreme@gmail.com> Cc: Greg KH <greg@kroah.com> Acked-by:
Paulo Marques <pmarques@grupopie.com> Cc: "Randy.Dunlap" <rdunlap@xenotime.net> Signed-off-by:
-
Ahmed S. Darwish authored
Use ARRAY_SIZE macro already defined in linux/kernel.h Signed-off-by:
Ahmed S. Darwish <darwish.07@gmail.com> Cc: Miles Bader <uclinux-v850@lsi.nec.co.jp> Signed-off-by:
-
Jeff Dike authored
This patch fixes some missing ptrace bits on x86_64. PTRACE_ARCH_PRCTL is hooked up and implemented. This required generalizing arch_prctl_skas slightly to take a task_struct to modify. Previously, it always operated on current. Reading and writing the debug registers is also enabled by un-ifdefing the code that implements that. It turns out that x86_64 is identical to i386, so the same code can be used. Signed-off-by:
Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by:
-
Jeff Dike authored
x86_64 needs some TLS fixes. What was missing was remembering the child thread id during clone and stuffing it into the child during each context switch. The %fs value is stored separately in the thread structure since the host controls what effect it has on the actual register file. The host also needs to store it in its own thread struct, so we need the value kept outside the register file. arch_prctl_skas was fixed to call PTRACE_ARCH_PRCTL appropriately. There is some saving and restoring of registers in the ARCH_SET_* cases so that the correct set of registers are changed on the host and restored to the process when it runs again. Signed-off-by:
-
Jeff Dike authored
The startup code panics a lot if anything goes wrong early on. This is wrong for several reasons, like the kernel isn't running, so you can't really be calling into it yet, but the harm comes from useful error messages being trapped in the printk ring where no one will ever see them. This patch changes these panics to perror and printf in wrappers which also exit. Normal, informational, prints are also wrapped so that fflush(stdout) is called after each one. This is so the output appears in the correct sequence in the event of an error. Signed-off-by:
-
Jeff Dike authored
Comment the lack of locking of data that's set up once at boot time. Also fixed a couple of bogus printks. Signed-off-by:
-
Jeff Dike authored
Formatting fixes in the register handling code. Signed-off-by:
-
Jeff Dike authored
Comment the lack of locking of the elf data extracted from the ELF headers passed to UML. Signed-off-by:
-
Jeff Dike authored
Add an error message when two umids are put on the command line. umid.h is kind of pointless since it only declares one thing, and that is already declared in os.h. Commented the lack of locking of some data in os-Linux/umid.h. Signed-off-by:
-
Jeff Dike authored
Fix formatting in the sigio code. Signed-off-by:
-
Jeff Dike authored
Comment the use of a mysterious-looking lock. Signed-off-by:
-
Jeff Dike authored
Tidying the irq code - make a variable static activate_fd can call kmalloc directly since it's now kernel code added a no-locking comment fixed a style violation Signed-off-by:
-
Robert P. J. Day authored
Fix apparent typo, where CONFIG_64_BIT should read CONFIG_64BIT. Signed-off-by:
Robert P. J. Day <rpjday@mindspring.com> Cc: Jeff Dike <jdike@addtoit.com> Signed-off-by:
-
Jeff Dike authored
Fix a bunch of style violations in mem.c. Signed-off-by:
-
Jeff Dike authored
We shouldn't be using the os wrappers from os code - we can use libc directly. This patch replaces wrapper calls with libc calls. It turns out that os_sigio_async had only one caller, which was in startup.c, so that function is moved there and its name changed. Signed-off-by:
-
Jeff Dike authored
Some style fixes in startup.c. Signed-off-by:
-
Jeff Dike authored
Add a couple of comments about some non-locked data. Signed-off-by:
-
Jeff Dike authored
Locking comments and emacs comment removal in the low-level memory and temp file code. Signed-off-by:
-
Jeff Dike authored
Some small locking and formatting fixes in the ubd driver. Signed-off-by:
-
Jeff Dike authored
Replace global queue and lock with per-device queues and locks. Mostly a straightforward replacement of ubd_io_lock with dev->lock and ubd_queue with dev->queue. Complications - There was no way to get a request struct (and queue) from the structure sent to the io_thread, so a pointer to the request was added. This is needed in ubd_handler in order to kick do_ubd_request to process another request. Queue initialization is moved from ubd_init to ubd_add. Signed-off-by:
-
Jeff Dike authored
CONFIG_HOST_TASK_SIZE doesn't exist any more. Signed-off-by:
-
Jeff Dike authored
kstack_depth_to_print can be made const. Signed-off-by:
-
Jeff Dike authored
A bunch of the signal handlers can be made static. Signed-off-by:
-
Jeff Dike authored
syscall_index and next_syscall_index turn out not to be used. Signed-off-by:
-
